78 
CHEMISTRY: MENZIES AND WRIGHT Proc. N. A. S. 
mally below 300° C. If, as is true, the solvent must differ by at least 
150 degrees in boiling-point from the solute in order that the volatility 
of the latter shall not vitiate the results, then the use is indicated of such 
solvents as boil below about 100°, and these solvents are indeed those 
most commonly used. The low solubility of certain solutes at the freezing- 
points of the available solvents is a factor against cryoscopy. On the 
other hand, the precision attainable in cryoscopy has, in the past, been 
considerably greater than that realized in ebullioscopy, even with ap- 
paratus fortified against error by the superimposition of a multiplicity of 
paraphernalia. 
For these and other reasons, a very great deal of 
work has been carried out by very many investigators 
with a view to improving the ebullioscopic technique. 
One may but mention the two dozen papers by 
Beckmann and his collaborators; 1 and the work of 
Hite, 2 Orndorff and Cameron, 3 Jones, 4 Ross Innes, 5 
Meyer and Desamari, 6 Drucker, 7 Sakurai, 8 Lands- 
berger, 9 Walker and Lumsden, 10 McCoy, 12 Smits, 13 
Rijber, 14 Ludlam, 15 Erdmann and Unruh, 16 Lehner, 17 
_^ Turner. 18 Even a single publication by one writer — 
~" Hite — reports the trial of over 100 different forms of 
apparatus. In all these, and many other investiga- 
tions, the bulb of a mercurial thermometer has been 
submerged beneath the surface of the boiling solution, 
and much of the work has been directed to the avoid- 
ance of the consequent irregular superheating. In our 
opinion, the two most successful improvements made 
since 1892 have originated in this country, the intro- 
duction, namely, by Bigelow 19 of electric heating, and 
by Cottrell 20 of a "lift-pump" which pumps the solu- 
tion over the bulb of the thermometer, which no longer 
need be submerged in the solution, but is located above 
the solution in the vapor phase. When Cottrell 's 
device is employed, there remain outstanding the 
errors inherent in the Beckmann type of thermometer, 
the error due to uncertainty as to actual concen- 
tration, and the very serious error due to change of 
boiling-point caused by change of barometric pressure 
during the observations. This last may, in part at least, be overcome by 
the employment of somewhat cumbrous "manostats," 21 by especially 
constructed aneroid barometers, 6 or by the use of a second, duplicate, 
apparatus 22 operated simultaneously and containing another Beckmann 
thermometer immersed in the pure boiling solvent. 
In 1910, Menzies 23 had described a simple apparatus, independent of 
fig. 
